Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Indinavir (Crixivan; Merck & Co., Inc., Whitehouse Station, NJ; Fig. 1A) is a peptidomimetic HIV protease inhibitor approved by the Food and Drug Administration for treatment of acquired immunodeficiency syndrome. The metabolism of indinavir in several species has been reported (Balani et al., 1995; Lin et al., 1996). Seven prominent metabolites (M1-M7) have been identified in human urine and characterized (Balani et al., 1995). Except for a direct N-glucuronidation product (M1) that is specific to higher primates (Balani et al., 1995), in vitro studies in both human and rat liver microsomes indicate that all the other metabolites (M2-M6) are products of oxidative metabolic pathways mediated by cytochrome P450 isozyme IIIA (CYP3A) (Chiba et al., 1996, 1997). In vivo pharmacokinetic studies revealed that oxidative metabolism is the major route of elimination for indinavir in rat and human and the contribution of conjugation to elimination is minimal (<0.5% of dose) (Lin et al., 1995, 1996). In addition, all oxidative metabolites observed in vivo were also formed in NADPH-fortified liver and intestinal microsomes, with the N-dealkylated metabolite, M6 (Fig. 1B), being the predominant metabolite (Chiba et al., 1997).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Chemical structure of indinavir shown as free base (A) and the predominant metabolite M6 found in rat intestinal lumen and rat liver microsomes (B).

Since CYP3A represents about 30 and 70% of total cytochrome P450 activity in liver and intestine, respectively (Watkins et al., 1987; de Waziers et al., 1990; Paine et al., 1997), first-pass metabolism is projected to play a significant role in poor and variable bioavailability of some drugs that are substrates for CYP3A (Back and Rogers, 1987; Schwenk, 1988). Intestinal first-pass metabolism mediated by CYP3A has been shown to be clinically relevant for several drugs such as cyclosporin A (Hebert et al., 1992; Wu et al., 1995), midazolam (Paine et al., 1996), and possibly other CYP3A substrates, including the HIV protease inhibitor, saquinavir (Fitzsimmons and Collins, 1997). Chiba et al. (1997) showed that small intestinal microsomes are capable of metabolizing indinavir, but their calculations indicated only a minor contribution whereas the liver plays a much greater role in the first-pass metabolism of indinavir in both rat and human. Although microsomes provide in vitro metabolic profiles, an in vivo system including natural transport barriers and intact epithelial cells may be more representative of oral delivery limitations.

P-glycoprotein (P-gp), encoded by the human MDR1 and rodent mdr1a/1b genes, is constitutively expressed in the brush-border membrane of intestinal enterocytes and the canalicular membrane of hepatocytes and transports structurally and functionally diverse compounds (Ambudkar et al., 1999). Several lines of evidence indicate that P-gp plays a significant role in the oral absorption and excretion of some hydrophobic xenobiotics. P-gp limits the oral bioavailability of drugs such as paclitaxel (Taxol), talinolol, and digoxin, and it has been shown to excrete paclitaxel directly across the intestinal wall (Sparreboom et al., 1997; van Asperen et al., 1997). In humans, individual differences in P-gp expression in duodenal enterocytes are correlated with area under plasma concentration-time curves for digoxin (Hoffmeyer et al., 2000).

It has been postulated that mucosal transport by P-gp and CYP3A metabolism are functionally linked components of a xenobiotic detoxification system that limits the bioavailability of several drugs (Benet et al., 1999; Wacher et al., 2001). There is substantial overlap in substrate specificity between CYP3A and P-gp (Wacher et al., 1995; Kim et al., 1999), and several modulators/substrates of P-gp and CYP3A have been shown to coordinately up-regulate the expression of these proteins in vitro (Schuetz et al., 1996) as well as in vivo (Huang et al., 2001). Altered levels of these proteins could affect the oral absorption and pharmacokinetics of some drugs.

Indinavir is known to be a substrate of both CYP3A and P-gp. The purpose of this study was to detail the intestinal metabolism of indinavir in vivo and to study the effect of CYP3A and P-gp on the intestinal transport of indinavir in an animal model.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Indinavir capsules were purchased from the University of Michigan Hospital Pharmacy. Authentic indinavir used in HPLC and LC/MS assay development and validation was a gift from Merck & Co. Dexamethasone (DEX)-treated rat liver microsomes were purchased from Human Biologics International (Scottsdale, AZ). Ketamine-HCl was obtained from Fort Dodge Laboratories (Fort Dodge, IA), and Rompum (xylazine) from Bayer Corp. (Shawnee Mission, KS). The rat cytochrome P450 IIIA ECL Western blotting kit was purchased from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK). Radiolabeled compounds included [³H]indinavir (3.2 Ci/mmol) and [³H]vinblastine sulfate (9.1 Ci/mmol) from Moravek Biochemicals (Brea, CA), [¹⁴C]mannitol (316 mCi/mmol) from ICN Pharmaceuticals (Costa Mesa, CA), and [¹⁴C]diazepam (50-62 mCi/mmol) from Amersham Biosciences (Piscataway, NJ). The scintillation cocktail EcoLite(+) was produced by ICN. All other chemicals were obtained from Sigma-Aldrich (St. Louis, MO) and were used as received.

Animals. Pathogen-free, male Sprague-Dawley rats (Charles River Laboratories, Inc., Wilmington, MA) weighing 300 to 400 g were used in accordance with a protocol approved by the Institutional Review Board Committee on the Use and Care of Animals (University of Michigan, Ann Arbor, MI).

Extraction of Indinavir from Capsule Formulation. The contents of 200-mg indinavir capsules were emptied into an equal volume of 0.2 N NaOH and ethyl acetate and stirred well. Undissolved particles were filtered out through gravity filtration. Multiple liquid-liquid extraction was performed in a separatory funnel, using ethyl acetate as the extraction solvent while adding diluted NaOH to the aqueous phase. The extraction solvent from the previous multiple extraction procedure was pooled, and magnesium sulfate was added as a drying agent. The resulting ethyl acetate solution was filtered and allowed to evaporate to approximately 50 to 100 ml. A steam bath was then used to redissolve the crystals and allowed to cool down to room temperature and evaporate slowly overnight. Indinavir free base crystals were obtained with vacuum filtration and washing with ice-cold ethyl acetate.

In Situ Rat Intestinal Perfusion. Procedures for the perfusion studies were modified from the method previously reported (Fleisher et al., 1989). Animals were fasted overnight with water ad libitum prior to each experiment. Anesthesia was induced with an intramuscular injection of 100 mg/kg ketamine and 20 mg/kg xylazine mixture, followed by an intraperitoneal injection of 40 mg/kg sodium pentobarbital. A midline longitudinal incision was made. After the desired intestinal region of approximately 10 cm was identified, inlet and outlet glass cannulas were inserted and secured by ligation with silk suture. The inlet tubing was connected to a 30-ml syringe that was placed in an infusion pump (Harvard Apparatus, Holliston, MA). All animals, perfusion solutions, and the pumps were enclosed in a Plexiglas thermostatically controlled chamber set at 30°C. Intestinal effluent samples were collected into preweighed polypropylene vials every 10 min for up to 120 min. Each effluent fraction was reweighed after collection. Steady-state water and solute transport rates were established within 30 min after initiation of perfusion at the flow rate studied. Water transport was corrected by the gravimetric method. Relative drug loss from the perfusate was measured by HPLC or LC/MS.

Intestinal Ussing Chamber Studies. Procedures for the intestinal Ussing chamber studies were the same as previously reported (Jezyk et al., 1999). Briefly, small sections of excised rat intestine, 2.5 to 3 cm in length, were opened along the mesenteric border. Assembled diffusion chambers were placed in a 37°C heating block, connected to a 95% O₂/5% CO₂ airlift, and filled with 5 ml of 37°C buffer, pH 7.4. The diffusion chambers and the airlift/heating block assembly were purchased from Precision Instrument Design (Lake Tahoe, CA) and Costar (Cambridge, MA), respectively.

Microsome Incubations. The oxidative metabolism of indinavir was confirmed in incubation preparations consisting of NADPH and DEX-treated rat liver microsomes. The incubation mixture (final volume of 0.5 ml in 50 mM phosphate buffer, pH 7.4) consisted of 1 mM NADPH, 0.5 mg/ml DEX-treated rat liver microsomes, and various concentrations of indinavir (1-10 µM). After a 3-min preincubation at 37°C, the reaction was initiated by the addition of 25 µl of 20 mM NADPH stock solution. A zero time point sample of 100 µl was taken immediately after initiation of the reaction. The mixture was then incubated for 10 or 20 min. The reaction was terminated, and the zero time point sample was treated with twice the volume of ice-cold acetonitrile and centrifuged at 14,000 rpm for 5 min. Supernatant was transferred to a clean tube and dried under nitrogen. The residue was reconstituted with 100 or 200 µl of acetonitrile/1% formic acid (50:50) and centrifuged again at 14,000 rpm for 15 min before the supernatant was stored for LC/MS analysis.

Immunoblotting. Rat small intestinal enterocytes were scraped off immediately following completion of perfusion experiments. Samples were prepared and stored for protein analysis according to the previously published method (Lown et al., 1997).

Analytical Method. The HPLC system consisted of a Shimadzu GT-104 degasser, FCV-10AL low-pressure gradient flow control valve, and LC-10AT serial dual plunger pump (all from Shimadzu, Kyoto, Japan), a Waters (Milford, MA) 712 WISP autosampler, an Applied Biosystems (Foster City, CA) 783 UV detector, a Hewlett-Packard (Palo Alto, CA) 35900C A/D interface box with HPIB, and Hewlett-Packard ChemStation Software.

Statistics. Treatment differences were determined using Student's t test. Significant treatment differences were specified in accordance with a p value of 0.05. Paired two-tailed test was used only when regional permeability values were determined in the same animal in the same experimental run. Error bars in all graphs represent the standard error of the mean.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Regional permeability, calculated from the water-corrected decrease in outlet perfusate indinavir concentration compared with inlet perfusion concentration (eq. 1), was significantly higher in upper small intestine than in lower small intestine or colon (Fig. 2A). This comparison was made at a 10 µM indinavir inlet perfusion concentration, which is approximate to indinavir total solubility at perfusion pH 6.5. The perfusate assay showed a substantial second peak in the jejunum that was not observed in the other regions. This extra peak was suspected and later confirmed with LC/MS to be the dealkylation metabolite of indinavir, M6 (Fig. 1), the major metabolite formed in vivo via CYP3A (Balani et al., 1995). Appearance of M6 was region-specific and decreased along the gastrointestinal tract. Intestinal metabolism was most significant when indinavir solution was perfused in rat upper jejunum and the relative appearance of M6 decreased in the midjejunal region. M6 was not detectable in the ileum or colon (Fig. 2B).

View larger version (8K): [in this window] [in a new window]   Fig. 2.   A, effective permeabilities of indinavir (solid bar) and mannitol (open bar) in various regions of rat small intestine determined from in situ perfusion. Perfusion solution was pH 6.5 MES buffer containing 10 µM indinavir, 5 mM mannitol, and trace amounts of [14C]mannitol. n = 3 to 5, , significantly different from jejunum, p < 0.05. B, relative amount of predominant indinavir metabolite M6 (expressed as the percentage of M6 over total luminal loss of indinavir; for details see Materials and Methods) in various regions of rat intestine determined from in situ perfusion. Indinavir dosing solution of 10 µM at pH 6.5 was perfused through 10 cm of upper jejunum (cannulated immediately after ligament of Trietz), midjejunum (35 cm down from ligament of Trietz), ileum (immediately before cecum), and 5 cm of colon, respectively. n = 3 to 4, , significantly different from upper jejunum, p < 0.05. N.D., nondetectable.

This regional appearance pattern of M6 is consistent with the distribution of CYP3A in the intestinal mucosa of these same animals (Fig. 3). Perfusion of rat jejunum as a function of increasing indinavir concentration (high concentration of indinavir was achieved by slight acidification of the perfusing buffer by 0.2 pH units) led to decreases in the relative appearance of M6 that were accompanied by increasing indinavir apparent permeability (Table 1). At 10 µM indinavir concentrations, the relative appearance of M6 decreased in the presence of ketoconazole, a potent CYP3A inhibitor (Fig. 4). The apparent jejunal permeability of indinavir also decreased in the presence of ketoconazole.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Immunoblot of rat cytochrome P450 IIIA. Enterocytes of rats were harvested and stored as described under Materials and Methods. Analysis was done by Tris-glycine SDS-polyacrylamide gel electrophoresis (Novex, San Diego, CA) on a 12% precast gel and Western blot with rat cytochrome P450 IIIA ECL Western blotting kit (Amersham Biosciences) under constant voltage. Lanes 1 to 10 contained the following sample and amount: lane 1, ECL molecular weight markers (5 ug); lane 2, liver microsomes from pregnenolone-16-carbonitrile-treated rats (5 ug); lanes 3 and 4, upper jejunal sample (60 ug); lanes 5 and 6, midjejunal sample (60 ug); and lanes 7 to 10, ileal sample (60 ug).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Fraction metabolite (open circle with dashed line) and apparent permeability of 10 µM indinavir (solid circle with dotted line) in the absence and presence of 10 and 50 µM ketoconazole, respectively, determined in rat upper jejunal in situ perfusion. Fraction metabolite was calculated as the percentage of M6 over total luminal loss of parent drug, and the apparent permeability of indinavir was determined based on total loss of drug from lumen (for details see Materials and Methods).

Because indinavir is known to be both a CYP3A and P-gp substrate, concentration dependence and inhibition studies were performed in both the perfusion and Ussing chamber system to evaluate the potential contributions of indinavir metabolism and export to limiting drug absorption in the small intestine. The fact that luminal M6 appearance was not observed in the ileum allowed for an unambiguous assessment of indinavir export in rat lower small intestine. Ileal perfusion studies showed no uptake of indinavir at lower drug concentrations (data not shown), suggesting the involvement of an export process. Luminal concentration of indinavir around its solubility at perfusion pH 6.5 (10 µM) resulted in an ileal permeability value of 8.97 ± 1.46 × 10⁶ cm/s, comparable with that of mannitol (5.25 ± 0.49 × 10⁶ cm/s), which served as a low-permeability marker in the perfusion experiment. Bidirectional transport of indinavir in ileal tissues was demonstrated in the Ussing chamber system and compared with other marker compounds (Fig. 5). Mannitol and diazepam, used as low- and high-permeability markers, respectively, had efflux ratios close to unity based on comparable effective permeabilities evaluated from drug flux in either direction. The presence of verapamil, a P-gp inhibitor, did not alter the permeability of mannitol or diazepam in either direction (data not shown). In the same ileal Ussing chamber study, indinavir, similar to vinblastine, showed an efflux ratio considerably greater than unity, suggesting that indinavir is a substrate for P-gp-mediated mucosal export (Fig. 5). This bidirectional transport of indinavir could not be eliminated with probenecid, or procainamide, or a combination of gly-sar, cephalexin, and enalapril (Fig. 6A); however, the net flux of indinavir was significantly reduced to 63 and 61% of control with verapamil and cyclosporin A, respectively. Finally, it was demonstrated that at high inhibitor-to-drug ratios (200 µM verapamil and 0.1 µM indinavir), bidirectional transport of indinavir was completely abolished (Fig. 6B), indicating that verapamil could completely inhibit indinavir export in rat ileum. Expression of P-gp in the rat ileal mucosa was also confirmed by Western blot analysis (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Apical to basolateral (solid bar) and basolateral to apical (open bar) permeabilities of mannitol (5 mM), diazepam (100 µM), vinblastine (10 µM), and indinavir (10 µM) across rat ileal tissue determined from Ussing chamber, with apical pH of 6.5 and basolateral pH of 7.4. Each data point represents duplicate determinations from three animals. Efflux ratio was calculated as the ratio of basolateral to apical over apical to basolateral permeability.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   A, net flux of 10 µM indinavir across rat ileal tissue determined from Ussing chamber in the presence of various inhibitors relative to the net flux in the absence of any inhibitors (control). Inhibitors included probenecid (10 mM), a combination of peptide transporter substrates (10 mM gly-sar, 10 mM cephalexin, and 5 mM enalapril), procainamide (10 mM), verapamil (200 µM), and cyclosporin A (200 µM), respectively. Flux was measured up to 2 h under apical pH of 6.5 and basolateral pH of 7.4. Each data point represents duplicate determinations with three animals. , significantly different from control, p < 0.05. B, apical to basolateral (solid bar) and basolateral to apical (open bar) permeabilities of 10 µM indinavir in the absence of inhibitors (control) and varying concentrations of indinavir in the presence of 200 µM verapamil. Numbers inside the open bar represent efflux ratios, determined by basolateral to apical permeability over apical to basolateral permeability. ++, significantly different from control apical to basolateral permeability, p < 0.01. , significantly different from control basolateral to apical permeability, p < 0.001. , significantly different from control basolateral to apical permeability, p < 0.0001.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The rat has been widely used as an animal model for projection of human intestinal permeability and to study the mechanism of intestinal transport of drugs with a spectrum of physicochemical and biochemical properties (Chiou and Barve, 1998). Human perfusion studies demonstrated an excellent correlation between intestinal permeabilities of the two species for a variety of compounds (Amidon et al., 1995). Although the oral bioavailability of indinavir is much lower in the rat than in humans (Lin et al., 1996; Yeh et al., 1998), this difference has been mainly attributed to differences in hepatic metabolism between the two species (Chiba et al., 1997). Therefore, the rat was used as an animal model to study the intestinal transport characteristics of indinavir.

In this study, drastically different permeability values were observed in the rat jejunum versus ileum (Fig. 2A). This region-dependent absorption profile permitted an experimental focus on drug export and metabolic transport components. In the ileum, where there is no complication from intestinal metabolism, the interaction of P-gp and indinavir was studied to offer an explanation for the low-ileal permeability observed from intestinal perfusion. In the jejunum, where intestinal metabolism was most significant, the role of CYP3A metabolism of indinavir was studied and shown to enhance apical uptake and absorptive flux of indinavir.

Indinavir is a fairly lipophilic weak base as defined by a log P_{octanol/water} of 3.0 at pH 6.5, which would indicate favorable permeability; however, rat ileal permeability was much lower than projected based on pH partition theory. Previous studies in cell culture and P-gp knockout mice models had suggested the involvement of P-gp in limiting the oral absorption of HIV protease inhibitors, including indinavir (Kim et al., 1998). P-gp is expressed on the apical side of epithelial cells in the intestinal tract of a variety of species including humans and rats (Benet et al., 1999). The efflux activity for P-gp substrates was reported to be highest in the ileum of rat and human whereas moderately expressed in duodenum, jejunum, and colon (Makhey et al., 1998). In a more recent report, in rat, p-glycoprotein-mediated digoxin export was higher in duodenum and jejunum compared with colon, but ileum was not tested (Sababi et al., 2001). In the present study using rats, we demonstrated the involvement of P-gp in limiting the apical to basolateral transport of indinavir in the rat ileum based on demonstration of saturable permeability at high concentrations from perfusion experiments (data not shown), greater basolateral-to-apical flux across ileal tissue (Fig. 6B), and transport inhibition by P-gp inhibitors, verapamil and cyclosporin A (Fig. 6A), in Ussing chamber experiments. This study also showed that multidrug resistance protein, organic cation, and peptide transporters were unlikely to be involved in indinavir efflux (Fig. 6A).

In contrast to the ileum, indinavir permeability was significantly higher in the upper jejunum (Fig. 2). Further investigation revealed that this favorable jejunal permeability profile was the result of enhanced apical uptake of indinavir due to its intracellular intestinal metabolism. Based on recent results showing that the indinavir metabolite, M6, is also a P-gp substrate (Hochman et al., 2001), possible competition for P-gp export between indinavir parent drug and the M6 metabolite may also play a positive role in indinavir absorption in the upper small intestine.

Previous in vivo human plasma data, as well as in vitro metabolism work in human intestinal and liver microsome systems, identified a dealkylation product, M6, as a predominant species of metabolite generated by CYP3A (Chiba et al., 1997). The validity of using rat as an intestinal metabolism model was confirmed with the identification of M6 as the major metabolite from rat hepatic microsomes as well as in situ rat jejunal perfusion. Studies in a CYP3A-induced Caco-2 cell culture model had demonstrated unidirectional transport of M6 to the apical side (Hochman et al., 2000), which is mediated by P-gp (Hochman et al., 2001). Based on this information, the parameter F_met, fraction of metabolite (eq. 2), over luminal drug loss is utilized in this report as a measure of the extent of intestinal metabolism in rat jejunum. Intestinal metabolism was found to be region-dependent (Fig. 2B) and consistent with the decreasing expression levels of CYP3A protein along the gastrointestinal tract, demonstrated in this study (Fig. 3) as well as by other laboratories (Kolars et al., 1992).

The reduction in F_met with increasing luminal drug concentration projects a saturation of jejunal metabolism of indinavir (Table 1). The highest indinavir concentration tested (50 mM) was included, assuming that in vivo concentrations of indinavir could reach supersaturation in the intestinal lumen either as a result of bile-salt solubilization or variation in intestinal pH. As a low pK_a weak base, indinavir total solubility is high at typical gastric pH and much lower at small intestinal pH. The permeability of indinavir was enhanced as the luminal concentration of indinavir increased, an opposite trend from that of F_met. Given that permeability was calculated based on total drug loss from the intestinal lumen, the permeability trend may indicate that at high intracellular metabolite concentrations, M6 serves to minimize drug export via competition for P-gp. This is consistent with high passive permeability of the lipophilic drug and poor passive transport of the more polar metabolite (Sababi et al., 2001). Further studies performed at 10 µM indinavir showed that drug permeability followed the fractional metabolite formed (Fig. 4), i.e., inhibition of metabolism by the CYP3A inhibitor, ketoconazole, reduced the fractional metabolite as well as reducing drug permeability.

The regional dependence highlighted in this study suggests that indinavir is well absorbed in the upper small intestine but poorly absorbed in the lower small intestine. Despite high lipophilicity, ileal absorption is compromised by mucosal drug export. In the jejunum, intracellular metabolism will serve to promote mucosal uptake of indinavir by maintaining sink conditions inside the cell, thus enhancing the concentration driving force of the drug across the mucosal membrane. In addition, drug flux and absorption will be enhanced compared with the ileum since the generated metabolite will compete for drug export into the lumen.

The transport mechanism demonstrated in this study is in agreement with the pharmacokinetic profile of indinavir in rat and human. The early t_max of indinavir in both species, 0.5 h for rats (Lin et al., 1995) and 1 h for humans (Yeh et al., 1998), are consistent with a short elimination half-life and an early absorption window in the upper small intestine. The major excretory path for indinavir is through fecal elimination, and the majority of radioactivity found in feces is in the form of metabolites (Balani et al., 1996). In addition to biliary excretion, intestinal metabolism and export of metabolites into the lumen may also be responsible for the recovery of metabolite-associated radioactivity in feces. The fact that indinavir has an early narrow absorption window might also explain its negative food effect, reported both in healthy subjects (Yeh et al., 1998) and in HIV-infected patients (Carver et al., 1999). Under fasted state conditions, indinavir is highly soluble in the stomach and could reach supersaturated concentrations in the intestine. At this high concentration, intestinal metabolism is saturated and hepatic metabolism plays a dominant role in the overall first-pass clearance process, as previously predicted from in vitro microsomal studies (Chiba et al., 1997); however, under fed state or other conditions, which may decrease either the luminal concentration of indinavir or the rate of drug delivery to intestinal epithelia, the potential for saturating intestinal indinavir clearance might be reduced. Under these conditions, intestinal metabolism in the jejunum and export in the ileum may play a more significant role in limiting the oral absorption of indinavir. The pharmacokinetic ramifications related to differences in intestinal transport and metabolism of HIV protease inhibitors are the subject of a subsequent article.